Combined Cooling Effects Between Street Green Space Morphology And Surrounding Built Environment

In the context of global warming and urban population growth,it is highly important to use street greening to mitigate the urban heat island effect,which can improve the quality of street space and enhance people’s well-being.Numerous studies have shown the feasibility of using greenery to mitigate the urban heat island effect.However,current research on quantifying vegetation is still primarily based on plant physiology and lacks a quantitative method for morphology in the context of planning and design.In addition,few studies have focused on the differences in the cooling effects of vegetation in different built environments.Through a literature review,this study found that the 3D morphological differences of vegetation result in varying degrees of cooling effects,which are not entirely linear relationships.Furthermore,the combined effects of vegetation and the surrounding built environment need to be further investigated.Therefore,this study will provide theoretical guidance and practical application references for the creation of urban street green spaces oriented toward thermal comfort.Firstly,this study quantified the street green space morphology based on deep learning and computer vision,then predicted the land surface temperature by the street green space morphology and the surrounding built environment.Subsequently,the performance of XGBoost,Light GBM,and Cat Boost models was compared,among which Cat Boost had the best performance and was selected as the prediction model for subsequent research.Then,the Shapley method’s Partial Dependence Plot(PDP)was used to analyze the non-linear relationship between street green space morphology and land surface temperature.Finally,the Dependency Plot(DP)was used to analyze the combined cooling effect of street green space morphology and surrounding built environment,and this mathematical relationship is summarized as a theoretical law,which is combined with the realistic law summarized by the original data to summarize the design strategy of urban street greening space under the guidance of thermal comfort.The research results showed that through the feature importance ranking,vegetation geometric contour morphology,leaf size,leaf density,and leaf color contributed more to the prediction of surface temperature.In the partial dependence plot(PDP),the various street green space morphology indicators showed different nonlinear relationships with surface temperature,which is consistent with various viewpoints in existing research.Based on the research results of the combined cooling effect obtained from the DP,it was found that when the surrounding buildings of the street have diverse skylines or a high degree of building coverage,the canopy of street green space should have certain undulations,and shrubs and grasses should be planted appropriately below the trees.Secondly,the canopy porosity of the trees should be balanced and not too large or too small.When the green space cannot avoid being in the shadow of buildings,light-colored foliage vegetation is more effective in reducing temperature.Conversely,the canopy of roadside trees should be coherent and full,arranged neatly with a certain distance between canopies,prune excessively long branches,use plants with large leaves,and increase porosity through proper pruning.When the exposure of green environment around the street is high,it can accommodate more types of green space forms,and there is no special requirement on the size of leaves,so long as the porosity is kept small.However,the streets exposed to low green environment have more restrictions on the green space form,and its form should not be too simple and rigid,but should show some ups and downs.Finally,the above research results and rules were applied to six streets in the Genshan community of Hangzhou city.Design strategies were formulated through defect analysis,and the effect diagram of the renovated street space was simulated according to the design strategy.all effect diagrams were predicted by the model proposed in this study.the results showed that the surface temperature in the renovated space was low,which verified the applicability and accuracy of the research results.